This is Your Brain on 5G

The rapidly expanding release of 5G wireless communications networks has spurred renewed concerns regarding the interactions of these higher RF frequencies with human tissue. In this talk we examine the relationship between RF exposure levels and dynamic heating in in-vitro bovine brain tissue at both ends of the proposed 5G spectrum — 39GHz on the high side and 4GHz on the low side. We compare our measurements with existing 4G exposure effects at 1.9GHz, and derive accurate dT/dP (temperature versus absorbed power) curves at all of these wavelengths. We also compare the heating characteristics of brain tissue with a common gel simulant (Triton X) often employed at 1.9GHz and lower frequencies, and show that there is a very strong deviation in the response of the gel and the brain tissue at the higher 5G frequencies, thereby negating some of the advantages of this particular type of simulant when evaluating thermal transport effects in the mm-wave bands. Our dT/dP and SAR (specific absorption rate) plots from the brain tissue measurements at 1.9, 4, and 39GHz can be used to accurately predict the heating at depth over a very wide range of absorbed power levels and help to elucidate, and extend downwards, the traditional boundaries between widely designated thermal and non-thermal effect regimes. Our simple but highly accurate (<0.1C) thermal measurement test set up is believed to be one of the most sensitive demonstrated for in-vitro tissue, and we have been able to detect changes as small as 0.05C at depths of up to 25mm in both brain tissue and Triton gel over the range of frequencies investigated. Finally, we examine the impact of RF absorption depth on thermal diffusion at the three frequencies and show that at 39GHz the impact of RF heating is confined to a very small surface region in the tissue and can produce a temperature rise in the tissue of more than 60C with only 1W of incident power and over time scales of only a few minutes. The measurements are supported by finite difference time domain simulations showing in detail, the distribution of our RF source power with depth and surface area in the tissue. We also show the effects (actual measurements) of rapid pulsing of the power at the different RF frequencies (1 microsec, 1msec and 1 sec at 50% duty cycle) and of short (1 sec) and long (30 sec) single on/off RF cycles.